一种新的WEB社群挖掘方法

随着互联网信息的增长,WEB挖掘已经成为数据挖掘研究的热点之一。论文关注于WEB社群挖掘。WEB社群是由对于同一主题感兴趣的人或组织创作的WEB页的集合。传统的挖掘方法由于只利用超链信息,所得出的社群或者太小、或者太大而没有意义。论文结合WEB页的内容信息和超链信息给出了一种基于分类方法的WEB社群挖掘算法。在包含8282个网页的WEBKB数据集上进行了实验,实验结果表明,该文算法可以挖掘出较大而且有意义的社群。

A New Approach for Mining Web Community

With the increasing of information on Internet,WEB mining has become the focus of data mining.This paper investigates the problem of mining WEB community.A WEB community is a set of WEB page which is created by some people or organizations,which concern the same topic.Traditionally,WEB community is mined out only by hyper-link analysis.However,the WEB communities mined in this way are too small,or are not meaningful.This paper presents a novel algorithm,which takes advantage of both hyper-link and text in the WEB pages,to mine WEB community by SVM classification.The paper reports the experimental result conducted on WEBKB data collection,which contains 8 282 WEB pages.The result demonstrates that the proposed approach can mine out big meaningful communities.